In 2022, another batch of AI bulls left the big factory

In 2022, the exodus of Internet scientists continues.

Domestic "big bull departure wave"

Wang Gang, head of the autonomous driving laboratory of DAMO Academy, left to start a business

In January 2022, according to Tech Planet, Wang Gang, vice president of Alibaba and head of the autonomous driving laboratory of DAMO Academy, left. According to people familiar with the matter, after Wang Gang leaves, he will choose to start a business in the field of cleaning robots and has already obtained financing.

Ali expressed his gratitude to Wang Gang for his contribution to the unmanned driving technology of the DAMO Academy and blessed Wang Gang.

According to public information, Wang Gang was a tenured professor at Nanyang Technological University in Singapore. In 2017, he joined Alibaba and has served as Chief Scientist of the AI Lab, Head of the DAMO Autonomous Driving Lab, Vice President of Alibaba Group, and Head of the DAMO Autonomous Driving Lab.

During his tenure as Chief Scientist of Alibaba Artificial Intelligence Lab, Wang Gang was the first-generation technical leader of Tmall Genie, and many of Gang's research results were applied in the Tmall Genie X1 released in July 2017.

In 2017, Wang Gang became the head of the autonomous driving laboratory of Ali Damo Academy. In September 2021, Wang Gang announced that the L4 autonomous driving product self-developed by DAMO Academy and the terminal logistics unmanned vehicle "Xiaomandonkey" have landed in 22 provinces across the country, with a cumulative delivery order of more than 1 million orders and express delivery for more than 200,000 people. And left the next 3 years of development goals - after 3 years, the size of the Xiaoman donkey fleet will reach 10,000, and the average daily delivery of 1 million parcels. Wang Gang said that the unmanned truck "Big Mandonkey" for open roads has also started research and development. This is the first time Wang Gang has analyzed Alibaba's autonomous driving strategic layout.

In 2017, DAMO Academy was established, and the autonomous driving laboratory undertook the task of integrating and unifying the autonomous driving technology of Alibaba Group, and Wang Gang led the team to form a research and development force and begin application-oriented technology exploration. Wang Gang stressed that Ali will continue to invest in autonomous driving for a long time.

Due to disagreements with the company, Cambrian CTO Liang Jun resigned

On the evening of March 14, 2022, Cambrian issued an announcement that Mr. Liang Jun, a core technician, notified the company to terminate the labor contract due to differences with the company. In fact, as early as mid-January, Liang Jun had already submitted an application for resignation to the company. According to the provisions of the "Labor Contract" signed by Liang Jun and the company and the relevant provisions of the "Labor Contract Law", the company recently went through the relevant resignation procedures for him; After his resignation, Mr. Liang Jun will no longer hold any position in the Company.

Cambrian said that Liang Jun's departure will have a certain impact on the company's research and development management. The company has established a complete R & D system, formed a professional R & D team, reserve a wealth of patented technology, Liang Jun's departure will not affect the company's technological innovation, will not have a significant adverse impact on the company's overall R & D strength.

According to public information, Liang Jun was born in 1976 and holds a master's degree in communication and information systems from the University of Science and Technology of China. From 2000 to 2003, he worked as an engineer at Huawei Technologies Co., Ltd. Beijing Research Institute. From 2003 to 2017, he worked in the Basic Business Department of Huawei Technologies Co., Ltd. and Shenzhen HiSilicon Semiconductor Co., Ltd., successively serving as engineer, senior engineer, chief engineer, technical expert, and senior technical expert. Liang Jun joined Cambrian in 2017, mainly engaged in R&D management before leaving, and served as deputy general manager and chief technology officer of the company, one of the company's core technical personnel.

Jin Rong, vice president of Alibaba and vice president of DAMO Academy, resigned

After Wang Gang, the Dharma Academy swept away the monks again.

In May 2022, according to Tech Planet, Jin Rong, vice president of Alibaba and vice president of DAMO Academy, left.

In 2014, Jin joined Alibaba and is responsible for Alibaba's iDST US R&D center in Silicon Valley. He has a long-term interest in statistical machine learning, focusing on big data analysis and its application in information retrieval, e-commerce and other fields, and his specific products include the "PailiTao" function in the mobile Taobao app. Later, he served as the vice president of DAMO Academy, responsible for the research field of machine intelligence, known as "super brain".

Jin Rong graduated from Tianjin University with a bachelor's degree, Peking University with a master's degree, and received a Ph.D. in computer science from Carnegie Mellon University. He was a tenured professor at Michigan State University, chairman of NIPS, SIGIR conferences and KDD committees, and a National Science Foundation Career Award. He has long focused on statistical machine learning, focusing on big data analysis and its application in information retrieval, e-commerce and other fields.

Mei Xiao, a Byte Intelligent voice technology engineer, left

In May 2022, Mei Xiao, a Byte Intelligent Voice technology engineer, left. In 2018, Mei Xiao joined ByteDance. Mei Xiao's AI-lab intelligent speech department at Byte is mainly responsible for artificial intelligence-related data production and business strategy formulation, mainly in the direction of speech synthesis technology.

According to public information, Mei Xiao has more than ten years of experience in the field of speech synthesis, and understands the application of linguistics in this technology and technical solutions to solve business problems. After graduating from the Graduate School of the Chinese Academy of Sciences, Mei Xiao worked as a senior researcher in the field of image and speech technology in the field of image and speech technology as a senior researcher, and then joined Tencent in 2016 as a senior engineer, responsible for the formulation of speech synthesis language libraries, synthesis algorithm evaluation, etc. In terms of business, Mei Xiao's team as a middle office provides speech synthesis technical support for a number of businesses, including Lv Bu, Liu Bei, Sun Shangxiang and other voices of King Glory Robot.

Chen Yilun, CTO of Huawei Autonomous Driving, left to join Zhang Yaqin's team

On July 13 last year, the Intelligent Industry Research Institute (AIR) of Tsinghua University announced that Chen Yilun joined Zhang Yaqin's team as the chief expert of intelligent robots of the Intelligent Industry Research Institute (AIR) of Tsinghua University.

Yilun Chen received his Ph.D. in Electronic Engineering from the University of Michigan, and received his bachelor's and master's degrees in electronic engineering from Tsinghua University. Prior to joining AIR, Chen Yilun served as CTO and Chief Scientist of Autonomous Driving System of Huawei's Intelligent Vehicle Solution Division, responsible for the design of Huawei's advanced autonomous driving technology solutions, and led the full-stack R&D of Huawei's first-generation autonomous driving system from 0 to 1. He also served as the chief engineer of machine vision of DJI, the vice president of Jingyan Intelligent Technology, and the technical expert and project leader of Eaton.

Dr. Xiao Lihua, former Vice President of Alibaba Group and President of Alibaba Cloud Research Institute, left to start his business

In May last year, Dr. Xiao Lihua (nicknamed "Xiao Bo"), former vice president of Alibaba Group, president of Alibaba Cloud's new retail business unit, and president of Alibaba Cloud Research Institute, left and founded Chi Heng Heyi Technology Co., Ltd.

It is reported that the company provides enterprises with one-stop digital intelligence transformation services, including consulting + software + operation and other services, that is, to help enterprises form a closed-loop of strategic transformation, business restructuring, organizational upgrade, IT/DT construction, digital intelligence operations, and can provide end-to-end solutions.

In more than 4 years of Alibaba, Xiao Lihua has been responsible for the Group's strategic key customer A100, Tmall Empowering Merchant New Retail (Project General PM), Alibaba Cloud Empowering Merchant Digital Intelligence Transformation, and the theory and practice of full-link digital intelligence transformation in various industries of government and enterprises.

Yang Hongxia, the technical leader of the pre-training large model M6 of the DAMO Academy, left

In October 2022, according to media reports, Yang Hongxia, the leader of the large model of Ali Damo Academy, left in early September.

Dr. Hongxia Yang is the technical leader of the very large-scale multimodal pre-training model M6. The M6 model focuses on multi-modality and multi-task capabilities, and its goal is to build the world's leading general-purpose artificial intelligence big model.

In March 2021, DAMO Academy released the first 100 billion parameter multimodal large model M6 in China, which attracted overseas attention. Jack Clark, former head of policy at OpenAI, commented publicly: "The scale and design of this model is amazing. This looks like a sign of the growing growth of many AI research organizations in China. ”

According to reports, Yang Hongxia's departure this time is due to personal family reasons. Tech Planet's report said that Yang Hongxia's departure is considered to be the adjustment of some commercial projects that are difficult to land. An Alibaba Cloud insider revealed that "many projects of DAMO Academy are very good from afar, and difficult to land at close range", although both are in cloud and technology, but DAMO Academy's projects are far away from business production and research, and they are rarely sold together with cloud services. Therefore, the degree of application and commercialization of each project of DAMO Academy is a mystery.

Previously, Ali has always stressed that DAMO Academy does not need to have profit pressure. However, in 2022, Internet companies generally reduced costs and increased efficiency, and DAMO Academy has also made many adjustments.

Yang Hongxia is an outstanding AI scientist in the field of AI. According to the data, Yang Hongxia graduated from Nankai University with a bachelor's degree in statistics in 2007. She then went to Duke University to pursue a Ph.D. in Statistical Science, under the tutelage of Professor David Dunson. Yang Hongxia has more than 40 top papers. Former IBM Watson researcher, Yahoo! Director Data Scientist, etc. She has led her team to win the Super AI Leader (SAIL Award), the highest award of the 2019 World Artificial Intelligence Conference, the 2022 Forbes China Top 50 Women in Science and Technology, and the second prize of the 2020 National Science and Technology Progress Award.

Wang Yilun, head of the AI algorithm for ideal car intelligent driving and chief scientist of AI

In October last year, Wang Yilun, head of the ideal car intelligent driving AI algorithm and chief scientist of AI, left. After leaving, Wang Yilun may be involved in entrepreneurship. At that time, as soon as the news came out, the relevant person in charge of Ideal Auto said, "This is a normal flow of personnel. ”

During the ideal period, Wang Yilun is mainly responsible for the research and development of AI algorithms such as intelligent driving perception and maps. According to 36Kr reports, some sources said that it was related to too much pressure on internal research and development. Li Auto's self-driving team is less than half of Xpeng's, and CEO Li Xiang's pace of expansion is more conservative. Another person close to the ideal said that the reason for Wang Yilun's departure may also be because the ideal self-driving team is difficult to expand, and insufficient job indicators have not been approved for cost considerations.

Wang Yilun once said in an interview that in the ideal, work has become almost the whole of life. Under high-intensity overtime, he gained 20 pounds in one year. Ideal CEO Li Xiang also said frankly: In Ideal, it is normal for everyone to work until 12 pm, especially the autonomous driving team, which works most of Saturday and Sunday.

The situation overseas is also not optimistic

Tesla AI and Autopilot team leader leaves

On July 13 last year, Andrej Karpathy, Tesla's head of artificial intelligence and head of the company's self-driving team, announced his departure on Twitter.

Karpathy posted on Twitter: "Over the past 5 years, as Tesla has been working hard toward its goals, it has been a difficult decision to say goodbye to one of the others. Along the way, the Autopilot system has evolved from the original lane keeping assist feature to city street assisted driving, and I hope that more talented people will join the Autopilot team in the future to continue this momentum. Karpathy added that he has no next steps at this time and wants to spend more time "revisiting my long-standing passion for AI, open source, and educational technology work."

Tesla CEO Elon Musk replied on Twitter, thanking Karpathy for his work at the company.

Before joining Tesla in 2017, Karpathy was an AI scientist at OpenAI, has an extensive background in AI-related fields, and was the creator of one of Stanford's most respected deep learning courses. After Musk acquired OpenAI, Karpathy naturally entered Tesla.

Since joining Tesla, Karpathy has managed a team of senior machine learning scientists and engineers focused on building computer vision systems to support Autopilot's advanced driver assistance systems, which closely coincides with his previous papers and work at OpenAI.

Karpathy's 5 years at Tesla are also the "cornerstone era" when Tesla popularized the four words "artificial intelligence" from cars to general applications.

There is speculation that Karpathy's departure may be due to the delay in the delivery of the Full Self-Driving (FSD) feature set promised by Tesla for many years, and frequent Autopilot-related traffic accidents.

Several AI executives at Meta left, and the AI team went through a reorganization

Since the beginning of this year, Meta has had several executives leave.

In April, Meta CTO Mike Schroepfer left. Chief Operating Officer Sheryl Sandberg, who has been with Meta for 14 years, announced on June 1 that she will be leaving this fall. A day after Sheryl Sandberg announced her departure, Meta's vice president of artificial intelligence, Jerome Pesenti, also announced her resignation. Immediately afterwards, David Mortenson, a senior engineering director in charge of data centers and core infrastructure at Meta for 11 years, also announced his departure. His direct report, Santosh Janardhan, now vice president of engineering, will take over the position at a later date.

Last year, Meta's AI team also underwent a massive restructuring, including: responsible AI organizations will join the Social Impact team; The product team is responsible for protecting platform users, improving recommendations, making content more relevant, and improving advertising and commercial services, which will be transferred to the product engineering team; The AI4AR team will join Reality Labs' XR team; The AI research team FAIR will be the new pillar of Reality Labs Research. Its mission and charter will remain unchanged to drive fundamental breakthroughs in AI through research excellence, open science, and broad collaboration. The FAIR team will continue to have a strong leadership team, including Joelle Pineau, Antoine Bordes and Yann LeCun.

Google's 26-bit AI bull is gone

According to a survey by Business Insider, in the past year, Alphabet alone has had 26 AI experts leave, including subsidiaries Google and DeepMind.

There are many AI experts who leave and start their own companies.

For example, Adept AI is a startup founded in April 2022 that focuses on creating a bot based on the same core technology as a text generation tool, with its CEO, CTO, and chief scientist all from Google. David Luan, ECO of Adept AI, is a former vice president of engineering at OpenAI's California Labs, who joined Google Brain, was one of the authors of GPT-2, PaLM, and participated in part of GPT-3; Ashish Vaswani is the chief scientist at Adept AI. Niki Parmar is CTO of Adept AI and former Google Brain researcher.

Mustafa Suleyman, vice president of AI products and policy at Google and founding member of the DeepMind AI Lab, founded Inflection AI, a general modeling startup, and served as its CEO and co-founder. The company raised $225 million in funding last May. At least four Google AI employees joined Inflexion AI last year.

Last year, several experts from DeepMind left to join startups. Last January, for example, Martin Schmid, Matej Moravcik, and Rudolf Kadlec left DeepMind to found EquiLibre Technologies, an AI startup focused on the stock market and cryptocurrencies. All three are developers of DeepMind's AI DeepStack; Rudolf Kadlec, a former senior researcher at DeepMind and co-author of Player of Games, is currently CTO and co-founder of EquiLibre Technologies.

Write at the end

In recent years, the departure of scientists from Internet factories has gradually become a trend, some of them have changed factories, continue to take root in the industry, some choose to embark on the road of entrepreneurship, and some return to academia from industry...

This wave of scientists leaving will undoubtedly have a certain impact on large factories and the development of the Internet, for large factories, it may face the challenge of losing talents, but from another level, large factories also send a lot of talents to the industry, and these top AI talents are still an important force to promote AI innovation. However, the exact impact geometry will take time to evaluate.